On the Creation of a Sensor Only 11 Atoms in Size—Robbie Elbertse –Delft University of Technology in the Netherlands

Richard Jacobs: Hello, this is Richard Jacobs with the Finding Genius podcast. I have Robbie Elbertse, he is a researcher at the University of Delft or Delft University in the Netherlands. He has developed a sensor that has only 11 atoms in its size which is amazing. We are going to talk about that, how he built it, how he conceived of it, and what it can be used for. So, Robbie, thanks for coming. 

Robbie Elbertse: Yes, it’s my pleasure.    

Richard Jacobs: Yeah, so is this your concept, or is this someone else’s concept that you are working to advance? How did this come about?  

Robbie Elbertse: So, it’s not my concept. The paper that I worked on which got published, it has me and someone else, David Coffey as the main authors and it was David’s idea. He has been, he was doing post-doc in the group and I’m currently doing a Ph.D. in our group. So, he has a little bit more of experience in this field of research and he actually came about the idea because he wanted to use some kind of sensor to measure something but that sensor didn’t exist yet, so it has to be conceived. 

Richard Jacobs: What does this sensor measure? 

Robbie Elbertse: So, what this sensor measure is so-called spin waves and a spin-wave is kind of a magnetic wave that moves across atoms. So, generally, when we think about a wave, you can, for example, think of a stone falling into the water and it creates these ripples that go from the stone. Those are waves in the water but there are also, for example, electromagnetic waves, lights but if you only take the magnetic part then you get magnetic waves. 

Richard Jacobs: So, I’m familiar with electromagnetism but what are magnetic waves? Where do they come from? How do they propagate? Tell me something about them. 

Robbie Elbertse: Right, so these atoms, they are in my case, iron atoms and they are iron atoms on copper nitrate and these atoms have a so-called spin. It’s because they have uncoupled electrons that orbit around the nucleus and these electrons provide a magnetic moment and so, this magnetic moment is what gives the iron its magnetic properties. So, each iron atom can have its own magnetic moments, and let’s imagine we have a chain of 5 iron atoms in a row, depending on how you orient them, these atoms or these magnetic moments can all point in the same direction. Let’s say, they all point up for North. 

So, if we have them in a chain then these atoms are so-called ferromagnetically coupled because they all point in the same direction. a wave would then be kind of like the wave in a football stadium where the first atom instead of pointing up, it starts pointing down for a second, and then it comes back up and then the second one goes down for a bit and then comes back up and then the third one goes down for a bit and back up. So, it’s kind of like a wave of these atoms pointing either up or down and moves across this chain.  

Richard Jacobs: So, does this happen only in magnetic materials? Does this happen in disordered magnetic materials? When does this happen and what are the consequences of it? 

Robbie Elbertse: So, what you need are atoms that have such a spin, such a magnetic moment. So, not all atoms have this but iron is a good example of where this does happen, specifically in our case. So, I described the situation of a chain which is obviously very ordered but if you just throw a bunch of iron atoms in a clump together, even there, the atoms will have a magnetic moment pointing in a certain direction. It will not be as neat where all of them point in the same direction, they can point in kind of random various directions or by coincidence, point in the same direction. still, in these situations, there can be a magnetic wave but they are super hard to understand because for one if it’s already a disordered bunch of atoms, it’s hard to understand that, and then either on top of that if they start waving in different directions and it can have multiple waves all happening at the same time. 

So, while it is possible to have this in many different scenarios and situations, we try to find very simple situations so that we can understand them from their basic principles. 

Richard Jacobs: Well, okay, you wanted to develop a sensor to determine what; the magnitude or the existence of the magnetic waves in various materials? 

Robbie Elbertse:  Yes. So, we do know that when we have this chain of iron atoms that there are magnetic waves because we have been able to see how they reflect within such a chain but what we didn’t really know is how far it would travel. so, for example, there is this chain of 5 atoms. If I started the first step and I first atom and I initiate a magnetic wave, will it reach atom number 3, will it reach atom number 5, what if I make the chain 100 atoms long, will it reach the end of the chain. So, we had some theoretical guesses about that but obviously, it’s good to verify whether your theory actually holds true and that’s why we made the sensor. So, the sensor is able to detect whether such a magnetic wave reaches the end of the chain. 

So, we didn’t really measure in the middle of the chain but still if we can verify whether it reaches the end of the chain according to a certain probability that corresponds very well with our theory. Now, we can say very well, it appears that our theory correctly describes or seems to describe whatever we expect is going to happen and we find exactly that. So, we use this sensor on a chain of about 3, 5, or 9 atoms and we’ve found that the expected chance of the spin-wave reaching the end of the chain based on our calculations, corresponded very well to what we actually measure. So, that’s really cool. 

Richard Jacobs: So, what happens? What kinds of materials do magnetic waves propagate in? Do they propagate in all materials, just not very far? I mean, what; if I look at various materials, do these magnetic waves constitute anything important about their material properties or it’s only a certain material? 

Robbie Elbertse: So, you need the material to have such a spin, so these uncoupled electrons and iron is the case. A lot of the so-called 3D materials such as iron, cobalt, nickel, they all have uncoupled electrons. So, with them, you can have these magnetic properties and so you can have these magnetic waves. Let’s say, hydrogen, what else has uncoupled but helium, for example, helium doesn’t have any uncoupled electrons. So, with helium, you wouldn’t be able to do it. so, yeah, in general, most materials would allow for it. The nice thing about iron is that it has 4 uncoupled electrons so it has a large moment of two. I guess, each electron adds half of a magnetic moment, and also, it’s very nice because it is very abundant. You can easily do experiments with it. 

Richard Jacobs: So, if an element is electrically conductive, does that mean it’s going to be magnetically conductive? I don’t know if it’s the right term. 

Robbie Elbertse: So, the fact that something is electrically conductive has to do with the band structure of the material, so that means how much energy does it cost for an electron to be excited to a so-called conduction band and all metals have this conduction band very close to the so-called Ferro-Me energy so close to rest as the basis states. The ground state, we call it but that does not necessarily mean that it has magnetic properties. So, for example, Lithium is, no, actually Lithium can have uncoupled, an uncoupled electron 

Richard Jacobs: What about copper? Copper is very conductive. Magnetically, what is the vibe? 

Robbie Elbertse: So, bulk copper does not have any magnetic properties but if you look at single copper atoms then the story is different because single copper atoms, I believe have one single uncoupled electron and so if you have a chain of copper atoms, you will actually still be able to induce these magnetic waves in it.  

Richard Jacobs: Okay, and now to the sensor. How does your sensor work? I guess it’s a huge deal that your sensor is so tiny. 

Robbie Elbertse: Yeah, so the sensor, we say it consists of 11 atoms which are 8 in the middle and then 3 on the side. So, these 8 in the middle, it’s the so-called bistable bits. So, what happens is these iron atoms have a coupling strength with their neighbor so these 8 atoms in the middle, they have an anti-ferro me chromatic coupling which means that if atom number one points up then atom number two wants to point down, 4 wants to go up and then 4 wants to go down, so up and down and up and down. That is one state that has the lowest energy but there is also a different state that has the lowest energy which is down up, down up. So, both these states, they are called State A and State B have the least amount of energy and that means that they are stable. 

So if you start your chain of eights in State A, it can easily stay in State A because it’s stable there anyway but if you are able to bring it to State B, it will just stay in State B. So, what we then did is we have this bistable bit of 8 atoms, and so on both sides, the left and the right, for example, we add a smaller chain. On the right,. We have a small chain of 3 that we use as a reset lead and on the left, we have our input lead which is the lead in which we make these spin waves. Since it’s very weakly coupled to its neighbors, it can still fell its neighbors but if both neighbors show a similar kind of energy towards the guy in the middle, the guy in the middle is happy to stay at either A or B, whatever it was. Now, the trick is that we bring it into State A because then if we make an excitation in this magnetic wave, spin-wave in our input lead, it’s actually going to change a little bit of the energy landscape to make State B more favorable. 

So, then it will go from State A to State B but that only happens if the spin wave was able to reach the end of the chain. If the spin wave was not able to reach the end of the chain, it cannot influence this bistable bit in the middle and so, it will remain in State A like it initially was. So, how we do our experiments is we bring our bistable bits in State A, we try to induce a spin-wave or a magnetic wave and we go back to our guy at 8 and say, hey are you in State A or State B? If it’s in State A, then that’s a suggestion that the spin wave didn’t reach the end of the chain and if it’s in B then it probably reached the end of the chain. So, we repeat this experiment hundreds of times to get statistics, for example, 60% of the time, it reaches the end of the chain or 80% of the time and has some responsiveness of this bistable bit on our input parameters and so we change our input parameters. 

We make more excitations, more spin waves or we start the spin waves in a longer chain or we start them further away. So, like that, we are able to both understand how our sensor works but also how our input leads or our spin chain works. 

Richard Jacobs: So, what’s the lower end threshold of; I mean what constitutes a wave? If spin propagates through 3 atoms, is that a wave, or how it behaves? How sensitive is this instrument? 

Robbie Elbertse: Yeah, so, I mean, on one hand, that’s kind of a question when is something a wave and when it is just someone going. For example, in a stadium, if the whole stadium moves their arms up, it’s obviously a wave. If it’s just one person doing it and no one around him is doing it, is it a wave? Is it the start of a wave or is it just a guy flinging his arms? We will, in either way call all of them a wave because it’s conceptually easier to just expand that from let’s say two infinite amount of atoms but what we’ve found is that these waves, they only last for like 10 picoseconds which is very short and since they have a velocity of about 50 meters per second, that means that we are only able to reach about 5 atoms long before the chain kind of decays or before that spin-wave decays. 

So, when we made a chain of 9, we only saw that the first couple of atoms that we excited close to the guys 8 were actually able to change the 8 from A to B whereas if we try it all the way at the end of the chain, basically it didn’t affect our sensor. 

Richard Jacobs: But where do the losses occur when you start the magnetic wave? What’s happening? 

Robbie Elbertse: That’s a very good question. So, I keep on talking about these iron atoms but these iron atoms are not floating in the air. These iron atoms are on top of a surface and the surface is a single monolayer or a single atomic layer of Copper Nitrites and that all is on a large copper-gold substrate and this copper nitrate layer is meant to de-couple our iron atoms from the metallic substrate, the copper-gold metal but it’s not a perfect isolator. On one hand, we don’t want a perfect isolator because if it’s not a perfect isolator then the measuring tool that we use, so-called scanning something, a microscope is actually not able to work at all. So, we don’t want our iron atoms to be fully isolated but we do want them to be isolated enough that the electrons that are just floating around and going back and forth in our metallic substrate don’t decay our wave. So, what happens is you’ve got this wave. 

In general, if these iron atoms are isolated in free space, the wave would go on infinitely long and it would bounce back and forth, back and forth until eternity. But because it’s on this metal substrate, electrons from the metal substrate see these iron atoms and they hop on the iron atoms and they bounce back and as they do, they catch a little bit of the energy of the magnetic wave and so, this energy of the magnetic wave gets lost to the so-called bath of electrons of the metal underneath. 

Richard Jacobs: So, nature. How strong are magnetic waves in materials? So, they kind of hop in and out of existence all across a monolithic piece of iron like it would. What are the consequences of them happening or not happening? What does it look like, again, if you have a magnetized piece of iron? So, what’s happening?  

Robbie Elbertse: Yeah, so in a magnetized piece of iron what is happening is you’ve got billions, trillions, even more atoms, all having magnetic moment pointing in the same direction. Even there, it’s possible to have magnetic waves on top of this like coherence state which would cause like, a little bit of noise in the magnetic properties. But I would say, in general, it’s a very small effect in these bulk materials which is why we downsized to the point where we are not in bulk material yet and then these magnetic waves do not have a small effect anymore. So, the energy of these magnetic waves is in the order of milli-electron volts which is very small. So, that means if I’ve got a bulk iron at room temperature, just the fact that the iron atoms are wiggling around so much and the electron atoms are wiggling around so much due to the temperature, that already destroys any kind of magnetic wave. 

It will still exist; it will still happen randomly but it will decay so fast that you won’t really notice it. so, in our experiments, what we did is we did everything at one Kelvin. So, quite cold. 

Richard Jacobs: So, is it more than the sensor you were able to build was incredibly tiny or was it the fact that you were able to measure magnetic waves of more importance or both?  

Robbie Elbertse: Yes, for me personally, what is more, important is we are able to measure these magnetic waves because what happens initially before we were able to build this detector is we were able to induce magnetic waves but we can only measure them at the same location as where we induce them. But if we want to know where do these magnetic waves travel. In a chain, we have shown that they do travel but if we want to build more complicated structures and see how they travel in them, we want to have some kind of sensor tell us how these magnetic waves propagate in these different structures. Since our current measuring tool has not been able to do that, we needed to build something new. The fact that it’s like 11 atoms large, to us, whether it was 11 or 15, it didn’t really matter. 11 was just a number that it happened to be. The fact that it’s so small is a necessity in some sense because all of the things that we are investigating in our lab are always going to be in the atomic scale. 

So, it’s always going to be between one and let’s say 20 atoms. So, then if you’ve got a very large sensor, that’s going to be an issue as well. So, you want your sensor to be approximately the same size as whatever you are going to measure. 

Richard Jacobs: Are there any quantum effects because the size is so small? 

Robbie Elbertse: Yeah, definitely. So, I mentioned that this guy at 8 can switch from State A to State B, right.  

Richard Jacobs: Yeah, there is such a thing as magnetic state tunneling, or is it in multiple states at once? 

Robbie Elbertse: Exactly. So we are actually making use of that fact. So, normally, this guy of 8 would normally need to get enough energy in order of let’s say 10 milli electron volts which is something that we don’t provide it and it is something that temperature also cannot provide it yet it is somehow able to grow from State A to State B and that is exactly using quantum mechanics. So, there is a small overlap as we call it between certain states, and using that and the fact that there is a small energy difference allows for this switching from State A to State B. if it wasn’t for quantum mechanics, it would just be stuck in State A forever. 

Richard Jacobs: Okay, I thought you were encouraging the switching. 

Robbie Elbertse:  Yes, we are encouraging a switch by changing the energy spectrum but we are not changing it so much. That’s using classical mechanics, it will be able to go from State A to State B. 

Richard Jacobs: Okay, so, you are relying on the quantum mechanics that for a small percentage of the time, there’ll be this essentially this tunneling effect that causes a change in state by a complication. 

Robbie Elbertse: Yes, exactly. So, the so-called energy barrier between State A and State B is 10 milli electron volts but what we are doing is we are making a small difference between these two states in the order of 100 micro electron volts, so 100 times smaller. But even that small difference is going to be enough for it to want to go from State A to State B using this quantum mechanical effect of state tunneling. 

Richard Jacobs: So, the low temperature is probably important so that the wavelength essentially of all the atoms involved is big enough that it is overlapping and you have a coherent state, right? 

Robbie Elbertse: Yeah. So, we would not be able to do this at temperatures above 10 Kelvin. We have done the measurements at one and four Kelvin but if we start going higher then we run into troubles with the so-called thermo-sphering. 

Richard Jacobs: So, I mean most materials aren’t hanging out at one Kelvin. So, how is this going to translate to materials at room temperature or materials that are at any operating temperature commercially? What do you think is going to be the effect? Is there a long road ahead of you in terms of understanding these?    

Robbie Elbertse: So, I mean that kind of comes down to the question of what is the purpose of science. On the one hand, is it to understand the world, is it to develop the technology. I would say it’s a combination of both. You can only develop a technology once you understand the world better. In terms of whether our specific sensor is going to be, is going to end up in your phone at some point. I very highly doubt it. Instead, I would expect that this sensor can be used for us to understand better how these waves propagate. So, yeah, the goal of this was not so much to develop technology that is going to be commercialized, rather it was used to help the scientific community have extra tools in order for the scientific community to advance. 

Richard Jacobs: So, you don’t get any room temperature materials, regardless of what they are, what the effects are but it’s probably safe to say that in order to understand quantum effects and materials and certain states, a phase transition or variable temperatures or any state at all, you want to characterize it in quantum effects. One of those effects is this magnetic I guess conduction or induction, whatever you want to call it or propagation, is that a fair assessment? 

Robbie Elbertse: Yeah, in general, what science tries to do is just simplify everything until you can make all the numbers zeros, except for one. So, try to make temperature zero, try to make the number of atoms zero except for one then just try to simplify everything, and then start from the bottom up. Like that. 

Richard Jacobs: What do you think is happening in an integrated circuit? Now the gates are getting to be, I guess a couple of atoms thick. So, how do you think this magnetic wave propagation is affecting circuits? Maybe that’s the first use case in better understanding their properties and behaviors.  

Robbie Elbertse: Yeah, so, most integrated circuits right now are using still electric waves, so waves or electrons moving around. It has a disadvantage of also having heating because your electrons are moving and there is a little bit of fix in which causes heating. If all of these magnetic waves, there is no mass moving and sort of been heating. So, those are kind of still two different realms but it is true that as you make everything smaller, you are going to end up in situations where you are going to have spin waves regardless of whether we want them but you can still make these transistors using materials that do not show any magnetic properties and then you can kind of, overcome them. 

Richard Jacobs: So, okay. I guess, again, it’s a long characterization ahead to see what their significance is but at least you are getting there. This sensor, I mean making it itself, how did you get these 11 atoms isolated and then arranged properly? How fine-tuned are they? Are they truly in a line? If they are off by half an atom thick, how straight is this line? How good is it? 

Robbie Elbertse: Yeah, so, half an atom off would be very off for us. So, all of what we are doing is done in a so-called scanning tunneling microscope. So scanning tunneling microscope essentially is a super sharp needle, atomically sharp that we bring really close to our surface but not to the point that it touches. There has to be at least one atom in between. At this point, we have a different but kind of a similar tunneling effect where electrons tunnel from your super sharp needle to your surface, to your sample, to your substrate. This causes an electric current, so flow of electrons and it seems this flow of electrons or this occurrence is exponentially dependent on the heights basically, we can have our super sharp needle or tip to provide a kind of height map of our substrates. So, we can make images with that. That’s always nice. 

But we can also use this tip, this needle to literally pick up a single atom and place it somewhere else and since we are using a substrate that has square letters, we can basically end up putting our atoms on points anywhere on these square letters. So, it’s very easy to make straight lines just like that. 

Richard Jacobs: Again, how straight is straight? What’s your tolerance, do you think? If you are off by literally one atom’s width, can you tell and I would think that would make a huge change in your sensor whether it works or not.   

Robbie Elbertse: Oh, yes. So, what we do is our guy at 8, for example, there are two lattice sides in between the atoms and between the guy at 8 and the guy at 3 or the input chain, there are 3 latticed sites in between. So we are talking about a difference of 100 picometers or maybe 200 picometers and that is enough to significantly alter how they couple with each other. The good thing, however, is that the atoms like to stay in these latticed sides, and with our tip, we have the precision of about up to ten, perhaps even a single picometer. So, it’s not that difficult for us to pick it up and drop it exactly at a latticed site and even if we drop it slightly off-center, it will fall into the center. So, that makes it quite easy for us to align these things. 

Richard Jacobs: I wonder, could you use this as a ruler to judge the linearity of flatness of a surface?  

Robbie Elbertse: Yes, definitely. We are able to see the atomic steps. So, if you’ve got a surface that you think is atomically flat, give it to us and we can tell you how atomically flat it really is.  

Richard Jacobs: Yeah, I guess this will have a lot of uses, in addition to the magnetic wave. So, very interesting.  How long could you make this where it would literally just act as a ruler? If you made it, I don’t know, a hundred atoms long, do you think that would be very useful? 

Robbie Elbertse: I don’t know if it’s going to be useful as a ruler. At that point, you might as well just use your needle, move it across a certain distance, and using.  

Richard Jacobs: No, no. here is why I say it like when I see like electron micrographs, scanning electron or SEM,TEM, or whatever, what if you have, also, on the slide, again I was going to call it as a ruler or a certain number of atoms’ length, could you literally use some kind of machine vision to get you much more accurate sizes and dimensions of things when you are doing this? 

Robbie Elbertse: Yes, so there are some things called interactions with our group and different groups where they want something scanned using our scanning tunneling microscope and we can tell them how the surface looks like on it’s atomic details. 

Richard Jacobs: Okay, I guess it has a lot of uses which is great, you know. 

Robbie Elbertse: Yeah, so the scanning tunneling microscope was invented in 1973 and it got the Nobel prize, I don’t think it got it that year but shortly afterward.  

Richard Jacobs: Do they have any trouble resolving the size of things in these kinds of microscopes or like would your sensor work or help or there is really no need for it?  

Robbie Elbertse: so, I would say, the only limitation for this microscope to work is that whatever you are going to measure needs to conduct electricity. So, you cannot really measure anything that is non-conductive but anything else, you can measure up to atomic precision.   

Richard Jacobs: Okay, well, very good. Elbertse, what’s the best way for people to keep tabs on the progress of your work and sensor. How can they; where can they go? 

Robbie Elbertse: So, on one hand, there is our group website which is ottelab.eudelft.nl which is for my group but I also have a YouTube video that further explains how the system works. I’m trying to see if I can find the YouTube link and I can hopefully say it out loud. 

Richard Jacobs: Yeah, go ahead. Definitely, yeah. 

Robbie Elbertse: So, let me go to the YouTube page. So, the video would be youtube.com/isDElxuN64. Alternatively, you can just look up on YouTube and type for an 11 atom sensor and that should give you the video as well. 

Richard Jacobs: Okay, yeah, that’s a simple way to do it.  

Robbie Elbertse: Yes. 

Richard Jacobs: Okay, Robbie, thanks for coming on the podcast. I know it’s late for you. It receives and I appreciate you being here.    

Robbie Elbertse: I appreciate you reaching out to us. 

Richard Jacobs: Yeah.